Photochemistry of Ketone Polymers. II. Studies of Model Compounds

William R. Mateker , Thomas Heumueller , Rongrong Cheacharoen , I. T. Sachs-Quintana ... J. E. GUILLET , J. DHANRAJ , F. J. GOLEMBA , and G. H. HARTLE...
0 downloads 0 Views 514KB Size
KETONEPOLYMERS41 3

Vol. I, No. 5, Septeinber-October 1968

phasized that in the poly(t-butyl acrylate) system the trans form was found to photoisomerize to the cis isomer. The spatial conformation of the three-dirnensional trans isomer (gauche) of the ester is such that the t-butylmethyl groups are rotated out of the plane of the carbonyl group16 leaving the oxygen atom more susceptible to hydrogen bonding. The cis isomer cannot hydrogen bond so easily since the methyl groups stagger the carbonyl oxygen. It is apparent that bonding of an intermolecular nature is operative in these carbamate systems. The essential evidence lies in the fact that the (N-H O=C) bonds are broken upon dilution and the N-H absorb,snce ratio was temperature dependent in solution.'2a This is in marked contrast to the polyacrylate system. It is concluded that a dominant factor reducing the type I1 quantum yield in thin films is due to a releasing of the energy absorbed by the gauche isomer ria some nonradiative process o r processes. Qualitative evidence supporting a relationship between energy dissipation cia a nonradiative process and hydrogen bondirig was advanced by Cowgill17 in work on fluorescence quenching of tryptophanyl and tyrosyl peptides.

Conclusions The conclusions may be arrived a t by inspection of the data in Table 111. The influence of hydrogen bonding on the type I1 quantum yields from I was found to be dependent on whether the reaction took place in thin films or solution. In the solution phase experiments, 4 1decreased ~ from 0.16 in 3-methylpentane t o 0.04 in methyl alcohol. A solvent shift causing a n inversion of the n-T* and T-T* states was also fiound to occur. Thus, the variation in dII was explained o n the basis of a change in electron (16) J. M. Riveros and E. B. Wilson, Jr., J . Chern. Phys., 46, 4605 ( I 967). (17) R. W. Cougill, Biochitn. B ~ o p h j s Acfa, . 133,6 (1967).

TABLE111 COMPAR~SON OF TYPEI[ QUANTUM YIELDS FOR THE POLYCARBAMATE AND POLYACRYLATE AT 254-mp EXCITATION Compound

Medium

Polyacrylate Polycarbamate Polycarbamate Polycarbamate

Thin film Thin film Alcohol Hydrocarbon

Reactive state

911

n--R*

0.08 0.01

Pa*

0.04

ii-T

0.16

n-**

*

density distribution in the excited states. Comparisons of the electronic absorption spectra with the phosphorescence lifetimes gave evidence for this interpretation. Second, the significant lower quantum yield of I in film form compared to the polyacrylate film or its yield in alcohol was rationalized on the basis of degradation of the excitation energy ria some radiationless process occurring from a resonance-stabilized hydrogen bonded gauche isomer. Evidence for this was found from infrared studies which showed a photochemically inactive, intermolecularly bonded gauche isomer. In addition, in thin films was found to increase by a factor of 2 in the temperature region (>T,) where the concentration of N-H. .O=C bonds decreased by -1 0 %. Therefore, the photochemistry of poly(t-butyl Nvinylcarbamate) appears to be controlled by the subtle interplay of a t least two factors: (1) a solvent effect which changes the character of the lowest reactive state and (2) energy degradation cia hydrogen bonding. Further experiments that will hopefully clarify the mechanism will be the subjects of further investigations.

-

Acknowledgments. Dr. H. James Harwood of the University of Akron and Miss Anita VanLaeken of this laboratory are sincerely thanked for the syntheses of I and 11, respectively. Informative discussions with Drs. J. B. Flannery, D. L. Stockman, and M. S. Walker are acknowledged with pleasure.

Photochemistry of Ketone Polymers. 11. Studies of Model Compounds G . H. Hartley and J. E. Guillet Department of Chemistry, University of Toronfo, Toronto 5, On fario, Canada. Received June 7, 1968 ABSTRACT: Studies of the photolysis of model ketones of the general structure R(C=O)R in hydrocarbon solution show that the type I and type I1 quantum yields are functions of chain length. The type I quantum yield decreases rapidly, and reaches a limiting value of 0.012. The type I1 quantum yield decreases slowly and appears to be still decreasing in ketones containing about 40 carbon atoms, where it has a value of 0.06. Although the type I 1 process is insensitive to temperature and solvent viscosity, the type I process depends strongly on both, as might be expected from a process involving radical intermediates. The results of these studies indicate that the photochemistry ol'the largest of the model compounds is quite similar to that of the ketone polymers studied previously.

I

n a previous publication' we have described the photochemistry of polymeric ketones synthesized by the copolymeirization of ethylene and carbon monoxide. At a n early stage in the investigation it became ( 1 ) G. H , Hartley and J. E. Guillet, ,vacronzo/ecrt/es, 1, 165 (1968).

apparent that variations in the molecular weight of the polymer had very little effect on the quantum yields of the processes involved, and yet there was a marked difference between these yields and those for the relaOrganic molecules reported in the literature' tively It therefore became of interest to establish the molecular

414 G. H.HARTLEY AND J. E. GULLET

Mucromolecules

weight dependence through a study of model compounds. Accordingly a series of symmetrical aliphatic ketones of the general structure RC(=O)R were synthesized and studies made of the variation in quantum yield with changes in the length of the linear alkyl radical R. Absorption of ultraviolet light by aliphatic ketones leads to dissociation by the Norrish type I free radical process2 0

0

I1

I1

R-C-R--tR-C.

+ .R

(1)

At elevated temperatures, the acyl radical can decarbonylate, with the evolution of carbon monoxide. 0

II

A

RC,+

R.

+ CO

(2)

Ketones possessing hydrogen-bearing y-carbon atoms can undergo the Norrish type TI photoelimination to yield olefins and methyl ketones. 0

II

RCCHZCHZCHR’R”

hv

+RCOCHI + CHz=CR’R”

(3)

The type I1 reaction appears to proceed by intramolecular hydrogen transfer, yielding an enol which rearranges to the final ketone (eq 3a).4’5The hydrogen 0

It

RCCHsCHKHR‘R”

hv

+ OH

CH,=CR’R”

0

11 + RL=cH2 --f RCCH,

(3a)

transfer involves a six-membered cyclic in either a concerted or a stepwise process. In addition to the type I and type I1 reactions, a third primary process has been observed which results in the isomerization of the ketone to a cyclobutanol derivative.* However, this reaction has been studied in only a few ketones, and does not appear to be an important process at 3130 A, the wavelength used in the present work. The two reactions of most importance in polymer systems appear to be types I and 11.’ Recent quenching studies have confirmed that the type I1 process at least involves both the singlet and triplet excited states of the k e t ~ n e . ~ ,For ’ ~ the majority of ketones, however, the extent of the contribution of each state to the reaction is still unknown. In this paper we shall be primarily concerned with the effects of chain length, temperature, and solvent viscosity on the type I and type I1 reactions in linear, symmetrical, aliphatic ketones. (2) J. G. Calvert and I. N. Pitts, Jr., “Photochemistry,” John Wiley and Sons, Inc., New York, N. Y . , 1966, p 379. (3) C. H. Bamford and R . G. W. Norrish, J . Chem. Soc., 1504 (1935). (4) G. R. McMillan, J. G. Calvert, and J. N. Pitts, Jr., J . Amer. Chem. Soc., 86, 3602 (1964). (5) R. Srinivasan, ibid., 81, 5061 (1959). (6) F. 0. Rice and E. Teller, J . Chem. Phj.s., 6, 489 (1938). (7) W. Davis, Jr., and W. A. Noyes, Jr., J . Amer. Chenr. Soc., 60,2153 (1947). (8) N. C. Yang and D. D. H. Yang, ibid., 80, 2913 (1958). (9) P. J. Wagner and G. S. Hammond, ibid., 88, 1245 (1966). (10) T. J. Dougherty, ibid., 87, 4011 (1965).

Experimental Section Apparatus. The ketones were photolyzed with the 3130-A line of the mercury emission spectrum. The lamp used was a 250-W, medium-pressure (30 atm) compact source mercury arc, Type ME/D (Associated Electrical Industries Ltd.). The rest of the apparatus, including the filters and light monitoring system, have been described previously. Procedure. A quartz cell, path length 20 mm, was filled with 20 ml of pure solvent, and the intensity of the lamp measured at the temperature of the experiment. Enough ketone was then added to make a 0.1 M solution, which was mixed by a stream of nitrogen. Air was removed from the solution by alternate applications of vacuum and a slight positive pressure of nitrogen. The solution was then photolyzed under nitrogen, the cell being sealed by a Teflon stopcock. After photolysis (to about 5 % conversion), the ketone solution was replaced by pure solvent, and the intensity of the lamp checked. Analysis. After photolysis the ketone solutions were analyzed on a Perkin-Elmer Model 800 gas chromatograph, using dual flame ionization detectors and in. X 5 ft columns (5% w/w silicone SE30 on Chromosorb G . ) . The photolysis products were measured by comparing their peak areas with the peak areas obtained from standard solutions containing known concentrations of the products. The peak areas were measured by a disk integrating unit attached to the chromatograph recorder. Carbon monoxide evolution from the ketone photolyses was measured, in a separate series of experiments, as described previously. l Ultraviolet absorption spectra of the ketones and solvents were run on a Bausch and Lomb Spectronic 505 double beam spectrophotometer. Solvent viscosities were measured using a Desreux viscometer. Materials. The solvents used in the photolyses were purified by distillation and/or filtration through silica gel columns until they were completely transparent over the whole range of wavelengths passed by the filter system. Analyses were performed on the Perkin-Elmer Model 800 gas chromatograph. With the exception of 4-heptanone and 5-nonanone, all the ketones were prepared by the method of Davis and Shultz,” in which a carboxylic acid RCOOH is refluxed with iron powder. Decarboxylation of the resulting salt forms the ketone RC(=O)R in good yield and high purity. The ketones were either distilled or extracted from the reaction mixture and purified by redistillation or recrystallization. The 5-nonanone was Eastman Kodak White Label grade, and the 4-heptanone was Eastman Kodak Practical grade, distilled under vacuum through a dewar column. With the exception of 5-nonanone (purity 99.5 %), the minimum purity in each case was 99.9 mol %. Results and Discussion Dependence of Quantum Yield on Chain Length. The ketones studied were all aliphatic, linear and symmetrical. They ranged in size from a total chain length of 7 carbon atoms (4-heptanone), to a total chain length of 43 carbon atoms (22 -tritetracontanone). Figure 1 shows the effect of increasing chain length on the type I1 quantum yield for photolysis in paraffin oil solution at 70”. 411 appeared t o decrease quite regularly with increasing chain length, and was still as high as 0.06 in 22-tritetracontanone. Furthermore, even at chain lengths of this magnitude the type I1 quantum yield had not yet reached a limiting value and was still slowly decreasing. The value of 0.025 re(11) R. Davis and H. P. Schultz, J . Org. Chem., 27, 854 (1962).

KETONE POLYMERS 415

Vol. I , No. 5, September-October 1968

0 40

O 0 1I 24 l

I

I

I

I

J.

\

I

i 0°'/

t

0

0

I

30

20

40

11

CHAIN LENGTH TOTAL CHAIN LENGTH

Figure 2. Quantum yield for carbon monoxide evolution temperature, 120"; solvent, paraffin oil.

Figure I . Type II quantum yield ($11) as a function of chain length; temperature, 70"; solvent, paraffin oil.

($co) as a function of chain length;

ported for the type 11 quantum yield of a 1 mol ethylene-carbon monoxide copolymer is therefore of the expected order of magnitude for a very long chain ketone. 1 The above behavior can be contrasted with that of the type I reaction. Figure 2 shows the quantum yields of carbon monoxide evolution from photolyses of the ketones in parallin oil at 120". At this temperature the acyl radicals produced by the type I reaction should decompose quantitatively and the quantum yield for carbon monoxide evolution will be very nearly equal to that for the type I bond breaks. It can be seen that as the chain length of the ketone increases, #Joe quickly drops to a limiting value (0.012) and that the major decrease occurs between 4-heptanone (di-npropyl ketone), and 5-nonanone (di-n-butyl ketone). Calvert and Nico112 investigated the vapor phase photodecomposi1:ion of a number of propyl ketones. Their results show that at a temperature of 150", the substitution of one butyl group for a propyl group in di-n-propyl ketone lowers the carbon monoxide quantum yield by a factor of about 4, their values for 4heptanone and 4-octanone being 0.37 and 0.10, respectively. This was explained by Wagner and Hammond,13 who pointed out that abstraction of a secondary y-hydrogen atom, such as in 2-hexanone, should be easier than abstraction of a primary y-hydrogen, such as in 2-pentanone, with the result that the type I1 reaction should be faster than the type I. Their quenching experiments showed that such was the case.9 It is interesting to note therefore, that the type I1 quantum yield fix 4-octanone is higher than that for 4-heptanone, although 4-octanone has a lower total quantum yield for reaction. l 2 In the present work, it seems reasonable to expect that Snonanone, with its secondary y hydrogens, would have a higher type I1 rate constant than 4-heptanone. Figures 1 and 2 show, however, that the type I1 quantum yield for 5-nonanone is smaller than that for 4heptanone, and that the type I quantum yield is lower by a factor of about 16. This is presumably a result of an increased rate of radiationless deactivation in h o n a n o n e , compared to 4-heptanone.

From the above discussion we can provide a possible qualitative explanation for the difference in behavior between 4-heptanone and 5-nonanone. The type I and type II are probably competitive with each other and with the processes of radiationless deactivation. Energy dissipation by nonradiative processes occurs more rapidly in 5-nonanone, but since the type II reaction is also faster, the type II quantum yield is less affected than that for the type I reaction. Also, since the quantum yields continue to decrease with increasing chain length, reaction must become even less efficient in the long-chain ketones, compared with the processes of radiationless deactivation. Dependence of Quantum Yield on Temperature and Solvent Viscosity. The effect of temperature and solvent viscosity on quantum yield was investigated for one of the ketones, 8-pentadecanone (di-n-heptyl ketone). The type 11 reaction of this ketone produces 1-hexene and 2-nonanone.

(12) C. H. Nicol and J. G. Calvcrt, J . Org. Chem., 89, 1790 (1967). (13) P. J . Wagner and G. S . Hammond, J . Amer. Chem. Soc., 87, 4009 (1965).

0

CHI(CH2)6&H2),CHa

2

CH,=CH(CH,),CH3

0

11 + CH3C(CH2)eCH3 (4)

Table I shows the quantum yields obtained for these products from photolyses of 8-pentadecanone in paraffin oil solution. It can be seen that the type II reaction in 8-pentadecanone is not strongly temperature dependent. If we define an apparent activation energy by then a log plot (Figure 3) the equation #J = of the quantum yields in Table I gives a value of E of approximately 1 kcal/mol. This insensitivity to temperature change is a characteristic of the type I1 reaction in ketones which possess only one type of y hydrogen. Ethylene-carbon monoxide copolymers also show this temperature independence of 411.l TABLEI PHOTOLYSIS OF 8-PENTADECANONE TYPEI1 QUANTUM YIELDS Temp, 'C 30

60 90 120

145

-Quantum 1-Hexene

yields--. 2-Nonanone

0.066 0.068 0.091 0.087 0.092

0.067 0,067 0.086 0.091 0.102

416 G. H.HARTLEY AND J. E. GULLET

'

I '

t

"

'

I

1

"

Macromolecules

' ' 1 1

i

-1

i

7

i -1

1

I

25

I

31

29

27

I/T

I

x 10'

Figure 4. Log 4'" for 8-pentadecanone: '3, heptane solvent; 0, dodecane solvent; 0 , paraffin oil solvent.

l

25

,

/

, I/T

,

I

30

I

I

35

x IO'

Figure 3. Log (quantum yields) cs. 1/T for 8-pentadecaiione photolyzed in paraffin oil solution: 0, log $hexene; Q, log Pnonanone; 0 , log $heptane. At 90", a change in the solvent viscosity from 4.1 (paraffin oil) to 0.22 CP (heptane) produced no detectable change in quantum yield, indicating that the type I1 reaction in 8-pentadecanone is not viscosity dependent, The type I reaction in 8-pentadecanone produces heptyl radicals and carbon monoxide. The heptyl 0

/I C H ~ C H ~ ) B C ( C H ~hu_ )~CH~ CHa(CHd6.

0

il

4- .C(CHzhCH CO

(5)

+ .(CH2)eCH3

radicals can abstract hydrogen to form heptane. The quantum yield for heptane formation was measured in the photolysis of 8-pentadecanone in paraffin oil solution, and the results are shown in Table 11. The quantum yield increases rapidly with temperature. Furthermore, a log plot (Figure 3) of the values obtained is quite linear, the slope corresponding to an apparent activation energy of 4.8 kcal/mol. The carbon monoxide quantum yields for 8-pentadecanone in a number of solvents and at various temperatures are shown in Table 111. At moderate temperatures (60"), the quantum yield decreased with TABLE I1 PHOTOLYSIS OF 8-PEN'TADECANONE. QUANTUM YIELD FOR HEPTANE FORMATION Temp, 30 60 90 120 145

O C

$

x

102

0.22 0.42 0.87 1.4 1.9

increasing solvent viscosity, which would be expected if a cage mechanism was involved. The differences between the quantum yields were small, however, and of the same order as the probable experimental error, although the trend was perhaps significant. At higher temperatures the effect of solvent viscosity was even less pronounced, and a log plot of the quantum yields (Figure 4) shows that the solvent appeared to have little influence on the temperature dependence of the carbon monoxide quantum yield. The apparent activation energy calculated from Figure 4 was 5.0 kcal/mol. Conceivably, long chain hydrocarbon solvents could inhibit the type I reaction by the formation of a "cage" in which only a few segments of any one solvent molecule were involved. Under such conditions, beyond a certain limit, increases in the molecular complexity of the solvent would increase the viscosity, but would be virtually without influence on the quantum yield. Since the production of a molecule of carbon monoxide is accompanied by the formation of two heptyl radicals, then, if the only reaction of a heptyl radical was to abstract a hydrogen atom from the solvent, it would be expected that the quantum yield for heptane formation would be at least twice that for carbon monoxide production. An examination of Tables I1 and 111 shows that such is not the case and that &,, is generally of the same order as &,eptarle. The heptane product was measured by gas-liquid partition chromatography, and the silicone SE30 liquid phase used would not have resolved 1heptene and n-heptane. Consequently, disproportionation between a heptyl radical and another radical TABLE 111 PHOTOLYSIS OF ~-PENTADECANONE. CO QUANTUM YIELDS

Sol veil t

n-Heptane Dodecane Paraffin oil Dodecane Paraffin oil Dodecane Dodecane Paraffin oil

Temp, "C 60 60 60

80 90 loo 120 120

Solvent viscosity, CP

0.28 0.73 9.2 0.58 4.1 0.47 0.40 2.1

dco

x

10%

0.49 0.46 0.42 0.75 0.90 1.1 1.5 1.5

ALLYL-SUBSTITUTED CYCLOPENTADIENES 417

Vol. I , No. 5, Septeniber-October 1968

would not have changed the apparent heptane quantum yield. Other possible reactions of the heptyl radical are (a) recombination with another heptyl radical t o give tetradecane I